A measure of whether the performance difference between two ads or ad sets is actually real or just random noise. You need enough spend and conversions to reach statistical significance before you can trust test results and make decisions.
A measure of whether the performance difference between two ads or ad sets is actually real or just random noise. You need enough spend and conversions to reach statistical significance before you can trust test results and make decisions.
Rule of thumb: Most statisticians require 95% confidence and at least 100 conversions per variant before calling a test conclusive.
A measure of whether the performance difference between two ads or ad sets is actually real or just random noise. You need enough spend and conversions to reach statistical significance before you can trust test results and make decisions.